Effect of Nb Addition on Oxide Formation on Ti–xNb Alloys
2019 Volume 60 Issue 10 Pages 2204-2212
Details
2019 Volume 60 Issue 10 Pages 2204-2212
It had been reported that Ti–29Nb–13Ta–4.6Zr (TNTZ) alloy forms a dense oxide layer by high-temperature oxidation whereas CP Ti forms a multilayered oxide consisted of rutile monolayers and void layer. This morphological change is supposed to be mainly caused by Nb addition in Ti since the dense oxide layer of TNTZ consists of multiple oxide phases, at least with rutile TiO2 and TiNb2O7. In this study, high-temperature oxidation at 1273 K for 3.6 ks in the air of Ti–xNb alloys (x = 1, 5, 7, 10, 13, 15, 18, 20, 23, 26, 28, 30 and 32 mol%) was investigated to discuss the effect of Nb addition to Ti on its high-oxidation behavior, and on its oxide microstructure. From the results of the SEM observation, an oxide layer with a void layer was formed on Ti–xNb substrate from 1 mol%Nb up to 10 mol%Nb. However, densification of the oxide layer was confirmed at Ti–13Nb. Then, the dense oxide layer was formed up to 32 mol%Nb. XRD results indicated that only rutile-type TiO2 was identified from 1 mol%Nb up to 10 mol%Nb, then both TiO2 and TiNb2O7 were formed from 13 mol%Nb to 32 mol%Nb. These results indicate that dense oxide layer formation attributes to phase separation from TiO2 to TiNb2O7. Until 10 mol%Nb, the thickness of oxide layer was suppressed by Nb addition, whereas the layer thickness increased with increasing Nb content from 13 mol%Nb. The maximum exfoliation resistance of the oxide layer was obtained at 20 mol%Nb. The results of oxide growth rate at each Ti–xNb alloys suggested that Nb diffusion in Ti may rate-determining process of the dense oxide layer formation.
This Paper was Originally Published in Japanese in J. Japan Inst. Met. Mater. 82 (2018) 232–239.
Fig. 7 BSE images and Ti, Nb, O composition maps of Ti-1, 10, 13, 20Nb.
Metallic biomaterials, such as Ti alloy and Co–Cr–Mo alloy, are used in guide wires, artificial joints and, bone-fracture fixation parts,1–3) representing about 80% of implantable devices (implants). As a means of body-functional reconstruction,4,5) replacement or support for damaged body parts with prostheses made from biomaterials for defective parts, they can contribute to early functional reconstruction.6,7)
Ti and Ti-based alloys, which represent commercially pure titanium (commercially pure Ti, CP Ti), and Ti–6Al–4V, have excellent mechanical and non-magnetic properties, corrosion resistance, and biocompatibility. They are important, not only in the biomedical field, but also in the aerospace, automotive, and chemical industries. Ti and its alloys are also used in the dental field, due to their high specific strength and excellent corrosion resistance. For example, prosthetic teeth, such as dental implants (abutments, screws and fixtures), crowns, arch wires, and denture bases are the main applications in the dental field.8)
In recent years, there has been a growing demand for biomedical devices to restore, not only function, but also appearance (aesthetics), in terms of improving quality of life (QOL).9) In addition to this reason, a shortage of the number of dental technicians, and the popularization of CAD/CAM manufacturing systems for dental devices, promotes the use of ceramics and composite resin, as the development of these materials, with high esthetic and functional restorability, is achieved. On the other hand, due to superiority in processability, mechanical properties, and biostability, expectations on metallic materials remain high among clinical providers, such as dentists and dental technicians.
Miura-Fujiwara et al. reported on the coating techniques to improve the aesthetic properties of Ti alloys as a result of Ti oxide opaque coating, by conducting high-temperature oxidation treatment on CP Ti and Ti–29Nb–13Ta–4.6Zr (TNTZ), which is a biodegradable β Ti alloy.10–14) The oxide coating mainly consists of rutile TiO2, which is typically used as a white pigment and food additive. TNTZ is an alloy developed by Niinomi et al.15) to reduce the bone resorption caused by the stress-shielding effect.
In terms of the required mechanical properties of the coating for dental devices, the sufficient exfoliation resistance of the coating is important. In this regard, the oxide layer of TNTZ exhibit high exfoliation resistance, of 68 MPa on average, while that of CP Ti shows 2∼3 MPa on average in adhesive tests.16) The difference in the exfoliation resistances is caused by the microstructure of the oxide. In CP Ti, a multilayered TiO2 single-phase oxide layer, containing a void layer (gap), is formed10) and the exfoliation occurs mainly between these layers. On the other hand, the microstructure of the oxide layer of TNTZ is characterized as dense, fine, and uniform. The layer consists of rutile TiO2 and TiNb2O7 phases, with sub-micron grains. The major fracture mode of the oxide layer of TNTZ is cohesive failure near the interface between the metal substrate and the oxide coating.16) These results suggest that Nb addition to Ti alloy played an important role in the oxide’s microstructure characterization.
Additionally, it is well known that oxidation of Ti proceeds when the atmospheric pressure exceeds approximately 873 K, and the working temperature of fan blades and disks for aircraft engines is set at less than this temperature accordingly.17) There are reports on the effects of Nb addition on the high-temperature oxidation resistance of TiAl that has attracted attention as a material for light-weight, heat-resistant intermetallic compounds.18–20) In the literature, the effect of Nb addition at 0.2∼15 mol% is considered to decrease the thickness of oxide layer formed on the surface of TiAl alloy. The effect of Nb addition on oxidation resistance of TiAl is considered to be a result of the reduction of oxygen vacancy concentration by the principle of valence control. Therefore, in order to clarify the effects of additional elements on the oxidation behavior and oxide microstructure of Ti alloy, it is necessary to investigate the oxidation behavior of Ti–Nb binary alloys from dilute to high concentrations.
In this study, the effects of Nb addition on the high-temperature, oxide microstructure of Ti–xNb (x = 1∼32 mol%) binary alloys was investigated. The phase identification of the oxides, surface color evaluation, adhesive test, cross-sectional microstructure observation, and chemical composition mapping of the oxidized Ti–Nb alloy were investigated. From the obtained results, the effect of Nb addition on the high-temperature oxidation and its exfoliation resistance will be discussed.
Pure Ti, with a purity of 99.9% (Mitsuwa Chemical Co. Ltd., Japan), and Nb, with a purity of 99.5% (Nilaco Co., Japan), were used as base metals. They were weighed to obtain an alloy composition of Ti–xNb (x = 1, 5, 7, 10, 13, 15, 18, 20, 23, 26, 28, 30, and 32 mol%), and about 16 g per composition, respectively. An alloy ingot was prepared by arc melting. The melting was carried out in a high-purity Ar (≧ 99.995%) atmosphere at a pressure of 3.0 × 10−3 Pa or less. In order to make the composition uniform, the alloy was reversed 5 times per batch.
The obtained Ti–Nb alloy ingots were cut into 1 mm thick plates using a fine cutter, polished with emery paper up to #2000, and then subjected to wet polishing with 3 µm and 0.3 µm alumina powder abrasive, to make the surface uniform. The specimen was then subjected to ultrasonic cleaning in acetone for about 300 s to remove contamination on the surface caused by adhering substances, and then dried at room temperature. The mirror-polished specimen was subjected to an oxidation treatment in air, at a holding temperature of 1273 K and a holding time of 3.6 ks, after which it was cooled down to room temperature.
2.2 Phase identification by X-ray diffractionPhase identification of the surface of the Ti–xNb alloy after oxidation was performed by an X-ray diffractometer (XRD, RINT-2200 V, Rigaku, Japan), at Cu-Kα, 40 kV and 20 mA, with a diffraction angle of 2θ = 20∼60 deg. The phases were then identified from the diffraction peaks obtained by Powder Diffraction file.21)
2.3 Evaluation of exfoliation resistance by adhesive testingExfoliation stress was measured using an adhesion tester (ROMULUS, Quad Group Inc., USA). The schematic illustration of the machine is shown in Fig. 1. The testing set-up is as follows: first, the oxide surface of the specimen and the stud pin, having an ϕ2.7 mm adhesive surface, were glued with an epoxy adhesive. After the stud pin was fixed to the specimen, the specimen was placed downward on a cylindrical stage and the pin was connected to the main tensile jig, as shown in Fig. 1. Then the pin was pulled down with a loading speed of 102.97 N/s. A maximum load applied to the stud pin was obtained when exfoliation occurred. Exfoliation stress, σe, was determined from the maximum applied load divided by the adhesive area.
Schematic illustration of adhesion tester.
The preparation procedure for the cross-sectional samples is described below. As shown in Fig. 2, the specimen was cut with a diamond cutter, perpendicularly to the disc surface, along the disc diameter of the specimen. To protect the resulting oxide film, the specimen was embedded in resin, so that the cross-section could be exposed. After polishing to #2000 with emery paper, 3 µm and 0.3 µm of alumina powder abrasive and colloidal silica were used for buff polishing to obtain a mirror surface.
Illustrations of sample dimension for cross-sectional observation.
The cross-sectional microstructures of each Ti–Nb alloy after oxidation were observed by field-emission, scanning electron microscopy (FE-SEM, JSM-5310, JEOL, Japan). Oxide layer thickness was determined using secondary electron imaging (SE image). Elemental distributions of the oxide and the substrate were measured using an electron probe microanalyzer (EPMA, JXA-8600, JEOL, Japan). The elements measured were Ti, Nb, and O.
2.5 Depth-profile by GD-OESElemental analyses of Ti, Nb, O, and N in the depth direction, from the oxide surface to the substrate, were performed using glow-discharge, optical emission spectroscopy (GD-OES, GDA750, Rigaku, Japan). The measurement conditions were as follows: anode diameter of ϕ4 mm, voltage of 600 V, lamp pressure of 2.8 hPa, and sputtering time of 1.0 ks in an Ar atmosphere.
2.6 Activation energy evaluation of oxidationIn order to investigate the rate-determining process of oxidation during heat treatment, the activation energy E, for oxide growth, was calculated from the oxide layer thickness, using Arrhenius plot. From the diffusion distance equation $x = \sqrt{2Dt} $, the oxidation temperature was defined as T, the thickness of the oxide layer was defined as x, and the oxidation time was defined as t. The logarithm natural of D, ln D, was plotted on the vertical axis and 1/T was plotted on the horizontal axis. An approximate line was drawn on the obtained plot by the least-squares method, and the activation energy, E, was obtained from the slope.
Figure 3 shows XRD spectrum obtained from oxidized Ti–xNb alloys with various Nb contents, to investigate the effects of Nb addition on the oxide phase formed on the alloys after high-temperature treatment. When Nb content is less than 10 mol% (x < 10), only rutile-type TiO2 was identified. Weak TiNb2O7 peaks started to appear, in addition to rutile TiO2 peaks, in Ti–xNb alloys containing more than 13 mol%Nb, and clear TiNb2O7 peaks eventually appeared in Ti–18Nb. Therefore, it is suggested that TiNb2O7 precipitation occurred through the Nb addition of more than x = 13 mol% under the high-temperature oxidation conditions.
XRD profiles of Ti–xNb alloys investigated in this study.
In order to investigate volume fraction change of the oxides by Nb addition, peak intensities I of $(110)_{\text{TiO}_{2}}$ and $(110)_{\text{TiNb}_{2}\text{O}_{7}}$ in Fig. 3 were plotted against Nb content in Fig. 4. The peak intensity of $(110)_{\text{TiO}_{2}}$ decreased with increasing Nb content. In particular, in the range from 13 mol%Nb to 20 mol%Nb, the intensity of $(110)_{\text{TiO}_{2}}$ of 20 mol%Nb decreased by about 75%, compared with the peak intensity of 13 mol%Nb, while in the range from 20 mol%Nb to 32 mol%Nb, the peak intensity showed a gradual decrease of about 20% with increasing Nb content. On the other hand, the peak intensity of $(110)_{\text{TiNb}_{2}\text{O}_{7}}$ increased by 3.8 times with increasing Nb content, from 13 mol%Nb to 32 mol%Nb. The results in Fig. 4 indicate that the volume of TiNb2O7 phase in the oxide increased with increasing Nb content.
Nb content dependence on peak intensities of $(110)_{\text{TiO}_{2}}$ and $(110)_{\text{TiNb}_{2}\text{O}_{7}}$.
An adhesion strength test was conducted to evaluate the exfoliation resistance of the oxide layers. Figure 5 shows the relationship between the amount of Nb over 13 mol% and the exfoliation stress σe. In the case of 10 mol%Nb or less, the test could not be conducted because the oxide layer had fallen off before the test. For reference, the exfoliation stress of oxide coated CP Ti and TNTZ is shown in the figure.13) In the range from 13 mol%Nb to 18 mol%Nb in Fig. 5, the σe increased linearly, from about 5 MPa to about 20 MPa, with an increase in Nb content. The maximum stress was 72 MPa at 20 mol%Nb, which was almost the same as that of TNTZ (Ti:Nb = 7:2). After 23 mol%Nb, the exfoliation stress gradually decreased with increasing Nb content. At 32 mol%Nb, the exfoliation stress decreased to the same level as at 18 mol%Nb.
Relationships between exfoliation stress and Nb content.
In order to observe the microstructure of the oxidized specimen, a cross-section of Ti–xNb alloy (x = 1, 5, 10, 13, 15, 18, 20, and 26 mol%) was observed by SEM. Figure 6 shows the SE image of each composition. In the specimen of 1, 5, and 10 mol%Nb, which the oxide fell off before the test, voids of about 380 nm and ∼770 nm in diameter, which were continuous in the direction parallel to the oxide/substrate interface, were observed throughout the oxide layer, especially near the interface. Similar voids were observed in the oxide formed on CP Ti, and which formed a gap between the oxide monolayers.10) Similar to the fracture mode of the oxide on CP Ti, it is indicated that these voids were the origin of cracks and promoted the development of these cracks during the adhesion test. Therefore, it is suggested that the exfoliation stress became significantly lower than that of 20 mol%Nb. On the other hand, in the oxide microstructures in Ti–Nb alloys containing more than 13 mol%Nb, where TiNb2O7 phase formation in addition to TiO2 phase was observed by XRD, such voids or gaps were not observed, and a dense oxide layer, similar to oxide layer formed on TNTZ, was observed. In addition, the lamellar structure of the substrate, consisting of α and β phases, was observed through the interface to the oxide layer continuously, suggesting that the oxidation proceeded toward the substrate. In the cross-sectional SE image in Fig. 6, the lamellar structure is observed in the oxide layer near the interface in the alloys containing more than 15 mol%Nb. These microstructure changes suggest that oxidation occurred at the interface between oxide and substrate once oxidation started on the surface of the substrate, and the migration of the substrate/oxide interface occurred continuously toward the substrate of more than 15 mol%Nb at least. The above results indicate that the microstructure of the oxide changes from porous to dense, with phase separation by Nb addition, in a Nb concentration range from 10 mol% to 13 mol%, indicating that the oxide formation process changes in this concentration range.
Cross-sectional SE images of oxide layer formed on Ti–xNb alloys.
Elemental mapping of Ti, Nb, and O was performed by EPMA, to investigate the elemental distribution in the oxide layer. Figure 7 shows the back-scattered electron (BSE) image and elemental map of the cross-section of Ti–xNb (x = 1, 5, 13, and 20 mol%). The Nb-rich area is observed only near the oxide/substrate interface at 13 mol%Nb, which is consistent with the result that only a small number of TiNb2O7 diffraction peaks appeared at Ti–13Nb, as shown in Fig. 3. The Nb-enriched area appeared near the void in the image of Ti–1Nb. However, it is considered that they are caused by the edge effect, resulting from the unevenness because of the large void. In the case of Ti–20Nb, the Nb-enriched phase was widely distributed throughout the oxide layer.
BSE images and Ti, Nb, O composition maps of Ti-1, 10, 13, 20Nb.
Therefore, elemental mapping showed that the Nb-enriched phase was widely distributed throughout the oxide, and that the Nb-enriched phase formation contributed to the densification of the oxide layer and the improvement of exfoliation resistance, as shown in Fig. 6.
3.4 Relationships between Nb content and layer thicknessFigure 8 shows the relationship between the Nb content, x, and the oxide layer thickness, l. As shown in the figure, the oxide thickness decreased with an increase in Nb content, from 0 to around 13∼15 mol%Nb, and then the layer thickness increased with an increase in Nb content. Yoshihara et al.19) have investigated the effects of Nb addition on the high-temperature oxidation resistance of Ti–Al–Nb alloys, and reported that the layer thickness decreases with increasing Nb content, with up to 15 mol%Nb added, that is, the oxidation resistance was improved by Nb addition. The results from this study regarding oxide growth suppression effects of Nb addition at 13∼15 mol%Nb or less in Ti–Nb alloys are in good agreement with results reported by Yoshihara et al. The Nb content at which the layer thickening begins, almost coincides with the composition at which the phase separation of the oxide phase and densification of the microstructure start, suggesting that the phase separation by the addition of a certain amount of Nb is also related to the oxide growth rate.
Relationships between thickness of oxide layer and Nb content.
In order to clarify the rate-determining process of oxidation in Ti–Nb alloys, oxidation time dependence of layer thickness was investigated in Ti–xNb (x = 1, 5, 13, 20, and 26 mol%). In these experiments, holding temperature was 1273 K, and holding time was altered to 0.3, 0.9, 1.8, and 2.7 ks. Figure 9 shows the relationship between holding time, t, and layer thickness, l. From the results, t-l curves yield to parabolic law, as shown in Fig. 9. In addition, the layer growth rate was similar, regardless of Nb content between Ti–1Nb and Ti–5Nb, which forms a single-phase oxide, while the growth rate of Ti-13, 20, and 26Nb increased with increasing Nb content.
Relationships between thickness of oxide layer and duration time.
Arrhenius plots of the oxide layer growth in these alloys are shown in Fig. 10. Each composition can be plotted on the Arrhenius plot, indicating that the growth process of the oxide is a thermal activation process. Table 1 shows the activation energy, E, and diffusion constant, D0, calculated from the slope and interpolation obtained from the plot in Fig. 10, respectively. E at x = 1, 5, and 13 was close to the activation energy of O in Ti (EO = 196 kJ/mol),22) suggesting that the diffusion of O in Ti is the rate-determining process. Since E at x = 20 and 26 is close to the activation energy of Ti or Nb in Ti–Nb alloys (ETi = 247.1 kJ/mol, ENb = 258.4 kJ/mol),23) diffusion of Ti or Nb in Ti–Nb alloys is considered to be the rate-determining process. These results suggest that the change in the rate-determining process with increasing Nb content contributes to the densification of the oxide layer.
Arrhenius plot of oxide layer thickness on Ti–xNb (x = 1, 5, 13, 20, 26). Oxidation time: 3.6 ks, oxidation temperature: 1173∼1323 K.
Regarding effect of Nb addition on Ti oxide formed in TiAl–Nb alloys at high-temperature, Yoshihara et al. reported that Nb was first dissolved in TiO2, and then TiO2 and Nb2O5 were formed when Nb content exceeded the solid solubility limit in TiO2.19) According to the TiO2–Nb2O5 phase diagram,24) the solubility limit of Nb2O5 in TiO2 was TiO2 = 28.5 mol% and Nb2O5 = 2.1 mol% at 1273 K. When Nb content in the oxide exceeds the solubility limit, phase separation from TiO2 to TiO2 and TiNb2O7 occurs. Thus, the atomic ratio of Ti and Nb is Ti:Nb = 87:13. From the XRD profiles in Fig. 3, TiNb2O7 peaks began to appear from 13Nb, which agrees with the phase diagram. Therefore, it can be concluded that the Nb diffusion occurred during the high-temperature oxidation reaction in Ti alloys contained more than 13%Nb, since the Nb solubility limit of TiO2 was exceeded and, thus, the phase separation of TiO2 and TiNb2O7 occurred.
4.2 Effect of oxide structure on exfoliation resistanceThe results of the XRD profiles in Fig. 3, and the exfoliation stress in Fig. 5, indicate a suggestion that the amount of Nb addition relates to the exfoliation resistance of Ti–Nb alloys. At a Nb concentration of 10 mol% or less, where a TiO2 single-phase oxide layer was formed, the exfoliation resistance exhibited a low value, although layer growth was suppressed by Nb addition. The exfoliation stress of the CP Ti oxide layer was also low.6) However, the stress of the Ti–Nb alloy with a small amount of Nb, that is, less than 13 mol%Nb, investigated in this study was much lower than that of the CP Ti. On the other hand, in the case of the oxide layer of 13 mol%Nb or more, dense microstructure, consisting of TiO2 and TiNb2O7 phases, were formed and the exfoliation resistance drastically increased, up to about 68 MPa at 20 mol%Nb. The oxide layer thickness linearly increased with increasing Nb content up to 30 mol%Nb, as shown in Fig. 8, and exfoliation stress also decreased linearly, with layer thickness decreasing and Nb content increasing. The tendency of the relationships between layer thickness and exfoliation stress is qualitatively similar to exfoliation stress of the oxide layer formed on TNTZ.8)
Therefore, it is suggested that the improvement of the exfoliation resistance in Ti–Nb alloys containing more than 13 mol%Nb is attributed to the densification of oxide microstructure, due to phase separation by Nb addition beyond the solubility limit of TiO2. Thus, these results strongly suggest that high exfoliation resistance of the oxide formed on TNTZ is also derived from the same reason. The molar ratio of Ti to Nb, and that of Ti to (Nb + Ta) in the TNTZ, is 75:25 and 53:47, respectively, which is different to the ratio of Ti–20 mol%Nb in the present study, suggesting that Ta and Zr may also contribute to the exfoliation resistance to some extent.
Furthermore, the factors that influence the Nb addition on the exfoliation resistance of the coating include, not only the formed oxide phase and its structure, but also the substrate/oxide interface structure, the difference in the atomic volume between the substrate and the oxide, the difference in the thermal expansion coefficient, the layer thickness, and so on.14,24) It is assumed that the increasing of the layer thickness with Nb addition might be one of the major reasons for the decrease in exfoliation stress with increasing Nb content above 20 mol%Nb. However, the abrupt increase in exfoliation stress from 18 mol%Nb to 20 mol%Nb cannot be explained by the present results. TEM observation of the oxide/substrate interface will be required for further information.
4.3 Improvement of oxidation resistance of Ti–Nb alloysAt x ≦ 15, the oxidation resistance of Ti alloys was improved, that is, oxide layer thickness decreased through Nb addition to the alloy, as shown in Fig. 8. According to the literature,18–20,25,26) the following two reasons were indicated: (1) Oxygen diffuses through oxygen vacancies in TiO2. Nb (pentavalent) addition, which has a higher valence than Ti (tetravalent), reduces oxygen vacancies and, consequently, improves oxidation resistance by the principle of valence control. (2) A nitride layer of Ti formed near the oxide/substrate interface by Nb addition, which acts as a diffusion barrier and suppresses oxygen diffusion.
In order to confirm the presence of a nitride diffusion barrier, elemental depth-profile measurement by GD-OES was performed on the oxide layer formed on Ti-1, 5, 13 and 26 mol%Nb alloys. Figure 11 shows the relationship between the sputtering time, t, and the emission intensities, I, of Ti, Nb, O, and N. From the graphs, the broad N peak was detected where the intensity of Ti increased simultaneously, which is supposed to be near the interface. This result indicated that a nitride diffusion barrier was formed at the interface of oxidized Ti–Nb, as reported in the previous study.26) On the other hand, as shown in Fig. 8, Ti–26 mol%Nb is in the composition range where the oxide thickness increases with Nb content. Taking these results into account, it is considered that the decrease in the layer thickness, due to Nb addition below 15 mol%, where the TiO2 volume fraction is high, is also mainly a result of the effect of the decrease in oxygen vacancies due to Nb diffusion into the TiO2 in (1), rather than to the diffusion barrier caused by the nitride layer in (2).
Depth profiles of Ti, Nb, O, and N from oxide layer to substrate through oxide/metal interface measured by GD-OES.
Finally, the formation mechanism of the high-temperature, oxide layer of Ti–Nb alloy is discussed based on the results obtained in this study. Figure 12 shows a schematic illustration of the oxide formation process on Ti–xNb alloy. At the initial stage of oxidation, O coming from the atmosphere diffuses into the Ti–Nb substrate through the passivation film, since the solution solubility limit of O into Ti is high. Then, an oxide layer would start to grow at the substrate surface, or at the interface between oxide/substrate, where the solubility limit of O is exceeded.
Schematic illustration of oxide layer formation. (a) x ≦ 10, (b) 13 ≦ x.
As shown in Fig. 12(a), in the case of x ≦ 10, Ti atoms for oxidation are supplied by outward diffusion from the substrate into the oxide. When the diffusion of O is a rate-determining process for oxide growth, the TiO2 formation reaction at the surface is faster than the outward diffusion of Ti atoms to the surface. In this case, voids tend to be introduced underneath the oxide layer as a Ti-deficient layer. With the progress of oxidation, this oxidation reaction vacancy moves inward, instead of Ti outward diffusion. Therefore, it is considered that a Kirkendall void was formed at the oxide/substrate interface due to aggregation without disappearance of these voids, and a gap formed underneath oxide layer.16,27–31) By Nb addition, the reaction rate of the Nb-dissolved TiO2 layer might decrease and the diffusion rate of metallic atoms also decreases, resulting in a decrease in the growth rate of the oxide layer. On the other hand, when 13 ≦ x in Fig. 12(b), Ti and/or Nb diffusion is a rate-determining process for oxide growth in this range. The outward diffusion of Ti is suppressed with an increase in the amount of Nb with a smaller diffusion rate. In addition, excess Nb, exceeding the solubility limit to TiO2, precipitates as a TiNb2O7 phase during the oxidation reaction. Once a thick oxide layer is formed, oxidation proceeds at the interface of oxide/substrate, and O atoms may be supplied from the oxide layer to the substrate. That is, O inward diffusion become dominant in this Nb content range.
The rate of oxide formation decreases, since phase separation of the oxide involving the redistribution of Nb is required to occur at the reaction interface, thereby achieving a balance with the diffusion rate of metallic elements. Therefore, the formation of a Kirkendall void at the interface disappears, resulting in the densification of the oxide occurring.
From the results of the cross-sectional microstructure observation and activation energy evaluation of the layer growth, it is supposed that the oxidation reaction interface proceeds toward the substrate direction. However, since the oxidation reaction rate is relatively fast, the distance of Ti and Nb diffusion accompanying the oxidation becomes relatively short-range. Therefore, it is considered that the continuous interfacial microstructure with fine grains, which inherited the lamellar structure of the substrate, was obtained. It is suggested that the continuous interfacial microstructure is also one of the keys to improving exfoliation resistance of oxide layers formed on Ti–Nb alloys.
High-temperature oxidation behavior in the ambient atmosphere, at 1273 K of Ti–xNb alloys (x = 1, 5, 7, 10, 13, 15, 18, 20, 23, 26, 28, 30, and 32 mol%), was investigated to discuss the effects of Nb addition to Ti on its high-oxidation behavior, and oxide microstructure. The obtained conclusions are as follows:
This research was partly carried out with the funding of the Japan Society for the Promotion of Science JSPS (No. 16K06777), the Inter-University Cooperative Research Program of Cooperative Research and Development Center for Advanced Materials, and the Institute for Materials Research (IMR), Tohoku University (Proposal No. 18G0038). The authors are grateful to Prof. Kunio Yubuta and Ms. Akiko Nomura of IMR for their contribution to the fabrication of the Ti–Nb alloys, and to Mr. Yoshihiro Murakami, Ms. Kazuyo Omura, and Mr. Issei Narita of IMR for their contribution to the measurement of the EPMA and XPS experiments.
